“Differential risk of hospitalization among single virus infections causing influenza‐like illnesses”

Background Acute respiratory infections are a major cause of morbidity in children and are often caused by viruses. However, the relative severity of illness associated with different viruses is unclear. The objective of this study was to evaluate the risk of hospitalization from different viruses in children presenting with an influenza‐like illness (ILI). Methods Data from children 5 years old or younger participating in an ILI natural history study from April 2010 to March 2014 was analyzed. The adjusted odds ratio for hospitalization was estimated in children with infections caused by respiratory syncytial virus (RSV), metapneumovirus, bocavirus, parainfluenza viruses, rhinovirus/enterovirus, coronavirus, adenovirus, and influenza. Results A total of 1486 children (408 outpatients and 1078 inpatients) were included in this analysis. At least one virus was detected in 1227 (82.6%) patients. The most frequent viruses detected as single pathogens were RSV (n = 286), rhinovirus/enterovirus (n = 251), parainfluenza viruses (n = 104), and influenza A or B (n = 99). After controlling for potential confounders (age, sex, recruitment site, days from symptom onset to enrollment, and underlying illnesses), children with RSV and metapneumovirus infections showed a greater likelihood of hospitalization than those infected by parainfluenza viruses (OR 2.7 and 1.9, respectively), rhinovirus/enterovirus (OR 3.1 and 2.1, respectively), coronaviruses (OR 4.9 and 3.4, respectively), adenovirus (OR 4.6 and 3.2, respectively), and influenza (OR 6.3 and 4.4, respectively). Conclusions Children presenting with ILI caused by RSV and metapneumovirus were at greatest risk for hospitalization, while children with rhinovirus/enterovirus, parainfluenza, coronavirus, adenovirus, and influenza were at lower risk of hospitalization.


| INTRODUC TI ON
Acute respiratory tract infection (ARI) represents the most frequent cause of outpatient visits to health care systems and hospitalization. 1  Globally, the proportion of severe pneumonia episodes was 8.6% in 2000 and in 2010 an incidence of 11.5% was estimated for low-and middle-income countries. 6 In Mexico, children less than 5 years of age account for up to 26.8% and 25.5% of all ARI and pneumonia cases, respectively. 7 Many studies have sought to determine the viral etiology of ARI in children, and the importance of some viruses, such as influenza and respiratory syncytial virus (RSV), is well known. 8,9 Previous studies have detected one or more viruses in a high proportion of ILI cases, ranging from 76.6% of cases in physicians' offices, 1 65% in outpatients, [10][11][12] and between 39% and 60% in hospital settings. 11,13 In recent years, the sensitivity of diagnostic tests has improved and the number of detectable pathogens has increased. 14 However, in Mexico and other Latin American countries, there is limited information regarding the prevalence and detection of viruses other than RSV and influenza as a cause of severe infections. [14][15][16][17] In the present study, we analyzed children 5 18 For this subgroup analysis, we restricted the study to children who were 5 years old or younger with confirmed viral infections.

| Study variables
The primary outcome of interest was admission to the hospital as a marker of a severe ILI. Since the emergency departments (ED) at the participating hospitals were often utilized by patients seeking both emergency and nonemergency care, patients who were in the ED for <24 hours were considered as outpatients. All deaths, regardless of hospitalization status, were considered to have severe ILI. All patients had follow-up 14 days (by telephone call) and 28 days (clinic visit) after enrollment. Patients initially seen as outpatients who were later admitted to the hospital were included in the hospitalized group for this analysis.
Demographic information and medical histories were obtained from all eligible participants at the baseline study visit. A HHSN261200800001E. The content of this publication does not necessarily reflect the views or policies of the Department of Health and Human Services, or Westat, nor does mention of trade names, commercial products, or organizations imply endorsement by the US Government. controlling for potential confounders (age, sex, recruitment site, days from symptom onset to enrollment, and underlying illnesses), children with RSV and metapneumovirus infections showed a greater likelihood of hospitalization than those infected by parainfluenza viruses (OR 2.7 and 1.9, respectively), rhinovirus/enterovirus (OR 3.1 and 2.1, respectively), coronaviruses (OR 4.9 and 3.4, respectively), adenovirus (OR 4.6 and 3.2, respectively), and influenza (OR 6.3 and 4.4, respectively). bocavirus with primers specific for this virus. Because the aim of this study was to assess the severity of viral infections, patients in whom any of these bacterial pathogens were detected (either alone or in combination with viruses) were excluded from further analysis.

| Statistical analysis
After excluding those with bacterial infections, we compared potential risk factors for severe ILI among those who were hospitalized vs those who were treated as outpatients. Univariate comparisons were evaluated using simple descriptive statistics. Differences in the prevalence in categorical variables were compared using chi-square statistics, and differences in means in continuous variables were compared using T tests. Fisher's exact P-values were reported for any categorical variable with cell counts less than five.
Logistic regression models were used to compare relative ILI severity associated with a given virus. Hospitalization was considered    Table 2. Some viruses were detected as single infections more frequently than others; for instance, RSV and RV/EV were found as the single pathogen in most infections in which this virus was involved (78% and 68%, respectively), while bocavirus and adenovirus were more commonly found in the presence of other viruses (co-infection rate of 60% and 56%, respectively).

| RE SULTS
The majority of patients in this cohort were hospitalized.
However, there were notable differences of hospitalized proportions between the different viruses. Among those with single virus infections, the proportion of patients who were treated as outpatients or required hospitalization varied greatly depending on the underlying viral infection (Figure 1).

| D ISCUSS I ON
The advance in diagnostic methods with increased sensitivity to detect respiratory viruses in recent years requires continued research to understand the prevalence of these viruses in children and to compare their ability to cause life-threatening illness as single infections. Lower respiratory tract infections including pneumonia have increased rates that coincide when respiratory viruses are more prevalent 19 Consistent with previous studies, we found that RSV was highly associated with hospitalizations, appearing to cause worse ILI outcomes when compared to single PIV, RV/EV, coronavirus, adenovirus, and influenza infections. Metapneumovirus (hMPV), though less common than RSV (9.7% and 24.7%, respectively) was more prevalent in our study than others had previously reported (4-5.5%). 20 hMPV was significantly associated with hospitalization; however, its ORs were generally smaller than those of RSV. We also found that a delayed visit to the doctor was more common among patients that required hospitalization. This finding underscores the importance of educating the public about seeking medical care soon after symptom onset to prevent more severe ILI outcomes.